1. Field of the Invention
Exemplary embodiments of the present invention relate to an apparatus and method for recognizing the current position of a vehicle using an internal network of the vehicle and an image sensor; and, particularly, to an apparatus and method for recognizing the current position of a vehicle using information detected by a sensor equipped to the vehicle in cooperation with an internal network of the vehicle, and information received from an image sensor, rather than recognizing information about the current position of the vehicle using a signal sent from a GPS satellite.
2. Description of Related Art
The global positioning system (GPS) is widely utilized in various fields at present. In particular, the GPS is actively used as a navigation system for use in vehicles to guide a user to a destination, and is installed in camera and cellular phones to perform an essential role in the supply of various services.
However, in order to apply a new technology, such as lane change guidance or obstacle avoidance guidance, to vehicles, an appliance having accuracy higher than that of the existing GPS is required.
For example, accurate information about the current position is indispensible information for self-driving cars in the field of autonomous vehicles. For this reason, a high-performance GPS has been used in the field of the autonomous vehicle, but since the high-performance GPS is expensive, it has not become very popular. A low-performance GPS has a problem in that it is hard to use in the autonomous vehicle because of its error.
As a route searching method for reducing error in the position measured by the GPS and accurately estimating the current position, dead reckoning and map matching have been utilized.
Dead reckoning is a process of estimating the current position of a vehicle from a reference position by calculating the moving direction and the moving distance of the vehicle. However, since dead reckoning is subject to the accumulation of error, many factors such as speed and direction must be continuously corrected. There is a drawback in that if the initial position value from which dead reckoning is commenced is not accurate, the accuracy of dead reckoning is very low.
Map matching is a process of automatically identifying the road on which a vehicle is traveling using the linear relationship between GPS latitude and longitude coordinates and the traveling direction of the vehicle, and then appropriately processing a screen when the vehicle position on a road of a digital map is not consistent with the GPS coordinates. Korean Unexamined Patent Publication No. 10-2006-0132302 discloses a method for correcting the current position of a vehicle by extracting a specific object from an image signal taken by an image pickup unit and performing map matching on the extracted object and object information stored in a map database. However, the map matching for correcting the current position using an image sensor and the previously established map data has a drawback in that since image processing is required, the processing entails latency, so it is hard to use map matching in cases in which it must be processed in real time.